REST: Create embeddings
Creates an embedding vector representing the input text.
HTTP request
POST https://ai.api.cloud.yandex.net/v1/embeddings
Body parameters
Request schema: application/json
{
"input": "unknown",
"model": "unknown",
"encoding_format": "string",
"dimensions": "integer",
"user": "string"
}
|
Field |
Description |
|
input |
Any of string | string |
|
model |
unknown Required field. ID of the model to use. You can use the List models API to see all of your available models. of them. |
|
encoding_format |
enum Required field. The format to return the embeddings in. "base64" currently not supported, "float" required.
|
|
dimensions |
integer NOT SUPPORTED BY ALL MODELS. The number of dimensions the resulting output embeddings should have. |
|
user |
string [CURRENTLY NOT SUPPORTED] |
Response
HTTP Code: 200
OK
Response schema: application/json
{
"data": [
{
"index": "integer",
"embedding": [
"number"
],
"object": "string"
}
],
"model": "string",
"object": "string",
"usage": {
"prompt_tokens": "integer",
"total_tokens": "integer"
}
}
|
Field |
Description |
|
data[] |
Required field. Represents an embedding vector returned by embedding endpoint. |
|
model |
string Required field. The name of the model used to generate the embedding. |
|
object |
enum Required field. The object type, which is always "list".
|
|
usage |
Required field. The usage information for the request. |
Embedding
Represents an embedding vector returned by embedding endpoint.
|
Field |
Description |
|
index |
integer Required field. The index of the embedding in the list of embeddings. |
|
embedding[] |
number (float) Required field. |
|
object |
enum Required field. The object type, which is always "embedding".
|
Usage
The usage information for the request.
|
Field |
Description |
|
prompt_tokens |
integer Required field. The number of tokens used by the prompt. |
|
total_tokens |
integer Required field. The total number of tokens used by the request. |